In December 2025, the World Health Organization did something it had never done before. It issued a guideline recommending medicines for the treatment of obesity. The guideline, covering liraglutide, semaglutide and tirzepatide, said GLP-1 therapies may be used for long-term treatment in adults with obesity, excluding pregnant women.
It was also unusually frank about who would benefit. On current trends, the WHO said, fewer than 10 per cent of the people who could benefit from these drugs are projected to have access by 2030.
Most commentary on that gap focuses on price. Price matters. But the harder problem is hidden in a phrase that is easy to read past: long-term treatment.
A drug you take for life is a different kind of policy
The WHO counts more than a billion people living with obesity, and 3.7 million deaths associated with it in 2024. Against that burden, a class of drugs that produces substantial, sustained weight loss is a genuine advance.
But the evidence so far suggests the benefit lasts only as long as the treatment does. In the extension of the STEP 1 trial of semaglutide, participants who stopped the drug regained much of the weight they had lost within a year. The WHO’s own recommendations were graded conditional, partly because of limited long-term data, cost and health system readiness.
That changes the arithmetic of equity. A course of antibiotics is a purchase. A GLP-1 prescription, as currently understood, is a subscription. A health system that starts a million people on treatment is committing to pay for a million people every month, potentially for decades. A country that can afford the first year but not the tenth has not expanded access. It has created a million future treatment interruptions.
The subscription arithmetic
It helps to run the numbers with deliberately round figures, because the shape matters more than the precise price. Imagine a public payer that can negotiate a generic GLP-1 at some fixed monthly cost per patient, call it C. It enrols 100,000 people in year one. Treatment is indefinite, and suppose a fifth of patients stop each year for side effects, cost-sharing or fatigue, while the system keeps enrolling a new 100,000 annually because demand does not go away.
In year one the bill is roughly 100,000 × 12C. By year five, even with attrition, the standing cohort has grown to well over 300,000 people, and the annual bill has more than tripled with no change in price. The programme’s cost is driven not by the price per dose but by the size of the stock of patients, and a stock only grows while inflow exceeds outflow. Halve the price and you buy time, not a ceiling. The cohort keeps growing until the number of people stopping each year matches the number starting, and in this example that equilibrium sits around half a million patients, five times the first year’s intake.
Now add the people who stop. In the arithmetic above, tens of thousands leave each year. If the trial evidence holds in routine practice, most of them regain weight, and a payer has spent a year or two of treatment for a benefit that fades. That spending is not wasted if it bought cardiovascular protection during the treatment period, but the case for public funding rests heavily on keeping people on the drug. So the true unit of cost is not a month of treatment. It is a durable outcome, and that requires the payer to finance continuation, monitoring and follow-up as a single package.
This is why budget officials should be wary of pilot programmes that look cheap. A one-year pilot of 5,000 patients reveals almost nothing about the fiscal commitment it implies, which is a growing population of patients who will expect the drug next year and the year after.
Generic entry will help, but not evenly
This year is the start of a real shift. Key patent and data protections on semaglutide are expiring in 2026 across several large markets, including Canada, China, India and Brazil. Canada’s data exclusivity lapsed in January. In Canada, pricing rules mean the price of an injectable must fall substantially once several generics are on the market. Analysts at IQVIA have noted that the markets affected cover roughly 40 per cent of the world’s population.
This will lower prices, probably sharply in some places. It will not by itself solve the equity problem, for three reasons.
First, secondary patents and litigation can slow entry even after the core patent expires, and the protections expire at different times in different countries.
Second, a cheap injectable still needs a health system around it: diagnosis, follow-up, management of side effects, cold chain and a way to judge who is benefiting and who is not. The WHO guideline explicitly says these medicines should be delivered within a chronic care model supported by a capable health system. Many primary care systems that struggle to keep hypertension patients on cheap tablets are not ready to manage injectable therapy at scale.
Third, cheaper drugs tend to reach the people who already have doctors first.
The hypertension comparison deserves more attention than it gets. Blood pressure medicines have been generic and cheap for decades, yet control rates in many countries remain low. The bottleneck was never the pill. It was diagnosis, follow-up visits, stock management and patients’ ability to keep coming back. GLP-1s add weekly injections, gastrointestinal side effects that drive people to stop, and in some formulations refrigeration. If a system cannot keep people on a tablet that costs little, the fall in semaglutide prices will mostly show up in private pharmacies.
The two-tier market is already here
The most likely near-term outcome is not “access” or “no access.” It is a split.
In wealthier settings, people with insurance or cash will take branded or generic GLP-1s, sometimes for modest weight loss, sometimes through online services with thin clinical oversight. In lower-income settings, and among poorer groups in rich countries, the people with severe obesity, diabetes and cardiovascular risk, who stand to gain the most, will be the least likely to get treatment.
There is a quieter signal of how the WHO itself is thinking about priority. In September 2025, before the obesity guideline, it added GLP-1 therapies to its Essential Medicines List for managing type 2 diabetes in high-risk groups. That is a narrower door than obesity treatment for everyone who qualifies, and it is probably the more realistic one for most public systems over the next few years.
That is the uncomfortable incentive. The market rewards volume among people who can pay. Public health value is concentrated among people who cannot. Without deliberate prioritisation, the drugs will flow toward the first group.
A second-order effect follows. As generics make cash-pay GLP-1s cheap in middle-income cities, the drugs will become normal among the urban middle class before public systems have written any rules. Those rules then get written under pressure from voters who already use the drugs, which tends to produce broad eligibility and thin funding rather than narrow eligibility and reliable funding. The order in which access arrives shapes the policy that follows.
The strongest counterargument
The case against this caution is serious. Obesity drives diabetes, heart disease, kidney disease and some cancers, and those conditions are themselves lifelong and expensive. If GLP-1s prevent enough heart attacks, dialysis and amputations, a lifetime subscription might pay for itself, and treating it as a runaway liability would be a false economy. Semaglutide’s cardiovascular outcome trial in people with established heart disease and overweight, SELECT, showed a reduction in major cardiovascular events, which strengthens that argument for the highest-risk patients.
That is where the argument holds: among people at high risk, where the avoided events are frequent and near-term. It weakens as eligibility widens. For a younger, lower-risk population, the avoided costs arrive decades later, are uncertain, and accrue partly to other budgets, while the drug bill arrives monthly. Payers discount future savings for good reasons, including the chance that patients will have stopped treatment long before the heart attack that the drug was supposed to prevent. Offsets are real, but they justify prioritisation by risk, not open access.
What a fair rollout would prioritise
The WHO points to pooled procurement, tiered pricing, market competition, local manufacturing and voluntary or compulsory licensing as tools to improve access. Those address supply. Health systems also need rules for demand.
For health ministries and payers, the first rule is to prioritise by risk, not by request. Public money should go first to people with severe obesity and serious complications, where benefit per dollar is highest. The second is to budget for continuation before initiation, modelling the growing stock of patients over five to ten years rather than the first year’s cohort. The third is to track stopping as carefully as starting. Discontinuation rates are the most honest measure of whether access is real, and very few systems publish them.
For clinicians, the WHO frames medication as an adjunct to behavioural and lifestyle support, not a replacement, and the chronic care model it describes needs people who will review patients, manage side effects and decide when the drug is not working. That is time in primary care, which is already scarce.
For hospital leaders, the practical question is capacity. If eligible patients arrive in large numbers, endocrinology and obesity clinics will not cope, and triage will default to whoever is most persistent. Protocols that let nurses and primary care teams manage stable patients will decide who gets seen.
For funders and founders, the opening is in the systems around the drug: cheap ways to identify high-risk patients, adherence support that works in low-income settings, and supply chains for injectables outside cold-chain-rich cities. A generic molecule is not a business moat. Keeping the right patient on it for five years might be.
And across all of them, the drug should not crowd out prevention. Food environments, taxes and urban design still determine how many people need treatment in the first place, and a finance ministry paying for GLP-1s has a newly concrete reason to care about them.
What to watch through 2027
Four things will show whether the equity problem is being solved or merely repriced. The first is the actual price path of generic semaglutide in Canada, India, Brazil and China during 2026, and how quickly secondary patent disputes are settled. The second is whether any public payer in a low- or middle-income country publishes an eligibility policy tied to risk, with a multi-year budget, rather than an open-ended announcement. The third is real-world persistence data: how many people are still on treatment after one and two years outside trials. The fourth is whether pooled procurement or licensing mechanisms named by the WHO produce a concrete deal. Any of these would be more meaningful than another headline about falling prices.
The takeaway
The GLP-1 debate has been framed as a question of price, and price will fall. The deeper question is whether health systems can make and keep a promise to treat a chronic condition for life, for the people who need it most, rather than for the people who ask first. A cheaper drug is necessary. A durable commitment is what equity actually requires.
Is your health system planning for the tenth year of GLP-1 treatment, or only the first? Tell us how it is being handled where you work.
